Lateral Xray marks finding with ESS C Bozza
Lateral X-ray marks finding with ESS C. Bozza, E. Carrara, L. S. Esposito and G. Sirri o o o Goal: implementation of the lateral mark finding in the ESS software First Test: Plate-to-plate intercalibration study Second Test: Start a scan back without intercalibration zones G. Sirri – INFN Bologna 1
Lateral X-ray mark X G. Sirri – INFN Bologna 2
Mark recognition Not possible to be done: o No extra camera with low magnification optics are available n o Only a small amount of the lines can be scanned The two strip are not always long enough to print a clear crossing point n NO NO Not possible to scan only one FOV per mark The approach is: o Three segments of each strip are scanned and lines are obtained by a linear fit. The coordinates of intersection point are the mark coordinates. G. Sirri – INFN Bologna Intersection point 3
Mark finding procedure This is the procedure currently under test. 4. Repeat 2, 3 for the 2 nd line 1. Set the corner 2. Scan this edge to find a first line segment 3. Scan further two line segment here 5. Repeat 2, 3, 4 for the second mark G. Sirri – INFN Bologna 6. Evaluate the intersection points and evaluate the transformation 4
How the transformation is evaluated? Rdet = A · Rstage +B Affine transf. of the plane are composed of scaling (reflections), rotations, and translations. Emphasizing this order, the components of a transformation can be encoded with the IFS formalism: r – scaling (reflection) in the x direction s – scaling (reflection) in the y direction theta - rotation of horizontal lines phi - rotation of vertical lines e - horizontal translation f - vertical translation With 2 mark coordinates: up to 4 independent parameter can be evaluated r=s , theta=phi , e , f G. Sirri – INFN Bologna (same scaling for both axis, rigid rotation, translation) 5
Mark finding reproducibility 5 measure of the mark coordinates without removing the emulsion plate Intersection. Mark 1: ( 121722. 1 , 3101. 5 ( 121722. 4 , 3099. 6 ( 121722. 4 , 3107. 0 ( 121722. 4 , 3104. 7 ( 121722. 3 , 3102. 8 ) ) ) Intersection. Mark 2: ( 121820. 2 , 97942. 4 ( 121818. 3 , 97937. 0 ( 121820. 3 , 97943. 1 ( 121822. 9 , 97938. 3 ( 121820. 3 , 97936. 6 ) ) ) The system is able to find the marks within 10 microns ( better for the X coordinate ) G. Sirri – INFN Bologna 6
First Test: plate-to-plate intercalibration study o o A first test has been done by scanning 5 * 1 cm² zones in two adjacent emulsion plates of a brick exposed to cosmic rays. The scanning was done at LNGS Scanning station with a vacuum system equipped with a vacuum channel larger than the emulsion size. Adhesive tape 3 1 emulsion 5 4 glass 2 Vacuum channel o The emulsion were attached to the adhesive tape In this setup the mark lines images are not affected by the vacuum channel image o Scanning done using the mark finding module in the acquisition sw o G. Sirri – INFN Bologna 7
First Test: plate-to-plate intercalibration study o o Tracks of the first plate are projected to the second one without extra intercalibration Pattern match results: ZONE ZONE ZONE 1 1 2 2 3 3 4 4 5 5 DX : DY : DX : DY : E: E: E: G. Sirri – INFN Bologna 100 51 51 105 86 86 91 91 M: M: M: -14. 0 33. 0 -7. 0 -3. 2 -3. 7 -7. 6 10. 5 13. 5 -6. 5 9. 4 RMS: RMS: RMS: 8. 8 8. 2 11. 6 9. 1 8. 0 10. 3 5. 6 5. 9 7. 0 6. 7 8
First Test: plate-to-plate intercalibration study - zone 1 G. Sirri – INFN Bologna 9
First Test: plate-to-plate intercalibration study - zone 2 G. Sirri – INFN Bologna 10
First Test: plate-to-plate intercalibration study - zone 3 G. Sirri – INFN Bologna 11
First Test: plate-to-plate intercalibration study - zone 4 G. Sirri – INFN Bologna 12
First Test: plate-to-plate intercalibration study - zone 5 G. Sirri – INFN Bologna 13
First Test: Evaluation of the best transformation After matching the track couples can be used to evaluate the intercalibration between the 2 plates. The results have to be compared with the Identity transformation which is the one supposed to be used without intercalibration zones. 406 matches found. QMAP – 6 parameters Aff 2 D : 0. 999894 -0. 000181 0. 000207 0. 999933 18. 65 -20. 81 (DET=0. 999827) r: 0. 999894 s: 0. 999933 theta: 0. 000207 phi: 0. 000182 e : 18. 65 f : -20. 81 QMAP - 3 parameters (Edb. Affine class implemented) Aff 2 D: 1. 000000 -0. 000199 1. 000000 12. 09 -23. 41 (DET=1. 000000) r: 1. 000000 s: 1. 000000 theta: 0. 000199 phi: 0. 000199 e : 12. 09 f : -23. 41 Scaling and Rotation are very small (< 1/10000). The translation error is about 20 micron. G. Sirri – INFN Bologna 14
Second Test: start a scan back o o o The scan back of brick 8199 with a predicted track has been started for test. This scanning has been done at Bologna using a vacuum system with a standard vacuum channel To avoid interference with the image of the vacuum channel the emulsion plate was placed at the edge of the channel with lateral mark outside the vacuum area. Vacuum channel G. Sirri – INFN Bologna 15
Second Test: scan back history (5 plates) o Scan Back done without computing infrastructure assistance. Linking and Prediction Finding has done manually POS TOL = 100 micron PL PPX 1 -2968655 PPY -3247649 PSX PSY -0. 405 0. 508 2 3 4 5 -3248286 -3249011 -3249659 -3250362 -0. 388 -0. 381 -0. 393 -2968124 -2967578 -2967082 -2966531 o o 0. 506 0. 499 0. 503 SLOPE TOL = 0. 03 + 0. 05*SLOPE FPX -2968628 -2968073 FPY -3247628 -3247629 -3248362 FSX -0. 388 -0. 374 -0. 381 FSY 0. 506 0. 519 0. 499 FP 23 20 33 DPX DPY -27 -21 -27 -20 -51 76 -2967042 -2966546 -2966564 -3249708 -3250378 -3250327 -0. 393 0. 503 -0. 382 0. 503 -0. 383 0. 492 23 24 16 -41 48 16 16 34 -34 Plate 3 : not found also for the scan back with intercalibration zones Plate 4 : skipped by the scan back with intercalibration zones because optical marks were printed on the wrong emulsion side. G. Sirri – INFN Bologna 16
Technical considerations o o The sw implementation is working, but it cannot be considered in the final state. For the version I’m testing: n “Automatic search of the First Mark” procedure not tested n The Mark Finding can fail if used above the vacuum channel n n n Automatic recovery strategy and manual set procedure are not implemented No autofocus Sometimes it recognize the (wrong) X-ray line placed ~1 mm far from each vertical strip. G. Sirri – INFN Bologna 17
How to deal with … o For these tests n n Use of large vacuum area plate and attach emulsion on an adhesive tape Use the standard but put the markers outside the vacuum area n Use precise mechanical reference on the glass plate n Put the exact emulsion sizes in the map string n n n o Set the upper-right corner manually Failures if marks are outside the vacuum area; set the Z with care. Open the log file and check manually if wrong strip is recognize The development is in progress. G. Sirri – INFN Bologna 18
Conclusions o The current implementation of the lateral mark finding in the ESS software has been tested and is working correctly (with some technical prescriptions). o Mark finding reproducibility is good o Plate-to-plate intercalibration of the order of 20 micron. o The development to have a “foolish proof” version is in progress G. Sirri – INFN Bologna 19
- Slides: 19